ALASKAN GLACIER STUDIES slope up which one could easily climb. In 1909 the margin was in the main an ice cliff, too steep for easy climbing, rent by a number of recently opened cracks, and broken by thrust faults. The extreme recency of this breaking was made evident by the angularity of the edges of these crevasses, which had not been exposed long enough for ablation to greatly modify the original angularity of breakage. A veneer of from 1 to 8 feet of moraine rested on the upper portion of this newly-exposed ice cliff, and in it occasional young willows were growing. It had the appearance of an ice cliff recently steepened, broken, and protruded through the moraine cover, the change being easily seen by comparison of photographs made in 1905 and 1909. As the ice cliff melted, the moraine from above was sliding down, and with it the willows that were thus undermined. These rock fragments with a few bushes, together with the debris that had fallen away from the steep ice face, had formed a small talus at the base of the ice cliff. For fully half a mile bowlders, clay, and other morainic d6bris were cascading down the steep ice face. From the new ice cliff many small, muddy streams were issuing and uniting, forming a muddy glacial stream of considerable proportions in the V-shaped valley between Hub-bard and Variegated Glaciers, where previously there was only a small stream, from the head of the valley, with trickling, clear-water tributaries from the Hubbard margin. This augmented drainage united to form a muddy pond back of the bowlder beach, through which it escaped in one fairly large channel. Where this pond stood, and between it and the glacier margin, annual and perennial plants and willow bushes throve in 1905 and 1906, but in 1909 many of these were submerged either in the icy water or in the mud that it was depositing. On one willow, around which we dug in the freshly-deposited mud, we found bark over a foot below the surface, from which new roots were starting. Practically all the willows partly submerged in mud were still growing, which is taken as evidence that the deposits which were overwhehning them had been there but a short time. Continued Advance in 1910. Between July, 1909, and June, 1910, the junior author found that forward movement of the northwest side of Hubbard Glacier (PI. XLV1H) exceeded that in the southeast, where some points actually receded, although at least two points on the south coast advanced during this period and the crevassing near the margin increased decidedly. The specific measurements quoted below are based upon a resurvey of the ice front originally plotted upon the plane table map in 1909 (Map 8). The photographs taken in 1910 also show the changes when compared with 1909 photographs from the same sites. From Osier Island (Photo. Station C, Map 3) it was clear that there had been an advance of the ice point just west of the center. From Gilbert Point (Photo. Station D, Map 3) a very slight retreat of the west edge of the glacier was shown, the ice point next it in the northwest part of the glacier had advanced, apparently between six and seven hundred feet, the ice point nearest midglacier was either stationary or showed very slight retreat, an adjacent f.liff (the one projecting farthest in 1909) retreated nearly 1000 feet, a point of black ice nearer the southeast side of the glacier was shortened almost 500 feet and large areas of black ice on the east side were diminished or had completely disappeared. From Haenke Island (Photo. Station A, Map 3) it was clear in 1910 that the northwest portion and the east central cliff had advanced slightly since 1906. This photograph station was not visited in 1909.